Design probes elicit expression rather than answers
Aliases: cultural probe · probes that elicit
What it is
A design probe—cultural probe in the tradition of Gaver, Dunne, and Pacenti—is a pack of tasks sent or handed over: postcards, diaries, maps, disposable cameras, unfinished sentences. Its purpose is not to have people answer questionnaire items but to elicit expression: to let participants, in their own media, speak fragments, wishes, and embarrassments that direct questions miss. “What do you do in the morning?” yields a normative answer; “photograph the alarm clock when you hate it” yields a clue that can be read. It is separate from a technology probe: there are almost no system logs here; the yield is work.
Why it happens
Much design-relevant experience will not fit an interview Q&A: private habits, feelings without names, vague images of the future. Probes make the task a half-open object, lowering the pressure to answer correctly and raising permission to use images, maps, and broken narrative. Ambiguity and a poetic register are not decoration; they stop life being folded into options. Returned material is not a requirement; it is a trace that can move a designer and be followed up. Rewrite the probe as a scale or a required field, and elicitation collapses into a survey: people start supplying socially desired answers.
Studying it
Gaver and colleagues’ original practice stressed posting, a stretch of unaided completion, and designers’ reading of returns, not coding into frequency tables. Later HCI has thematized probes more systematically, but material should still precede categories. The unit is a single piece and its situational account, not a mean score. Follow-ups should pursue the making of the piece, not drop the piece and re-ask a questionnaire. Record how a designer jumped from a photograph to an idea, or the inspiration cannot be audited. Do not use return rate as the only quality metric; silence and blank pages are data too.
Where it stops holding
Elicitation depends on people being willing to play. In highly structured, time-poor, or unequal settings, probes become extra homework and get a perfunctory pass. Across cultures, poetic tasks can be rude or untranslatable. In medical or legal settings that need auditable answers, probes cannot replace informed direct questions. Teams untrained in reading images and narrative will waste the material or project a prefabricated scheme onto it. Unlike technology probes, design probes almost never yield reproducible use counts.
Applying it
- Let each task elicit one kind of expression (a break in the day, a corner that should not be seen, a metaphor for a technology); do not stack a questionnaire on the same card.
- Say there is no correct answer, and show a sincere “bad” return rather than a pretty sample.
- In interpretation, write what the material shows in one column and what it made you think in another, so ideas are not written back as “the user said.”
- When you need counts or validation, open another method; do not add a scoring column to the probe.